With the rapid development of computer and network communications technologies, people exchange text information through networks more and more frequently. However, text information can reach user terminals such as users' mobile phones easily, and therefore is often used as an advertising means, and is even used by some lawbreakers to send junk information such as fraud or pornographic information, which severely affects and disturbs users' normal life.
Existing junk text information filtering technologies mainly use keyword filtering and black and whitelist filtering methods to filter junk text information. However, in a keyword-based method for filtering junk text information, the keywords can hardly cover all illegal information, and a malicious user can avoid the keywords by, for example, adding a large number of characters or interfering characters such as punctuation marks to the content of text information, so as to prevent the text information from being filtered. Moreover, in a filtering method on the basis of a blacklist and whitelist, because text information is processed according to users, if text information comes from a user in the whitelist, the text information is not filtered out even if content of the text information contains junk information.
Therefore, the existing junk text information filtering technologies have low filtering accuracy and are inefficient.